A new population-based retrospective validation study, using ICES data collected between 2007 and 2014 in Ontario, has found that health records can be used to track service use as well as the needs of individuals experiencing homelessness. Researchers found that the algorithms tested were able to accurately identify up to 35 per cent of individuals who experience homelessness in a given year.
The best performing algorithm estimated a large increase in homelessness prevalence between 2007 and 2016, with approximately one in 200 adults in Ontario experiencing homelessness in 2016.
“We do not know how many people experience homelessness yearly in Ontario, or in the country as a whole. Current methods rely on shelter data, or large point-in-time counts that require large investments, and are only done in specific communities,” explains Lucie Richard, Lawson researcher and lead author of the study. “Also, not all people who are homeless use shelters. This approach gives us an alternative that can include a wide range of communities and different types of homelessness over time.”
While numerous computer algorithms are already used to analyze administrative databases and identify disease populations to measure the burden of chronic diseases in Canada, this study, conducted by researchers at Lawson Health Research Institute and Western University, is one of the first to validate and track a social determinant of health with good accuracy.
People who experience homelessness often face mental and physical health challenges, have an increased risk of illness and death, and are among the highest users of emergency and other health care services. Homelessness is an important health issue for Canadians.
This ability to track individuals experiencing homelessness over time using administrative databases provides inexpensive, long-term monitoring, which will inform efforts to improve housing and health status in the community and across the country. Researchers and health care providers can now track health status and health care usage of individuals experiencing homelessness.
Understanding how many people experience homelessness and what demographics correlate with homelessness will allow for better service planning, and providers will have the information needed to more appropriately meet the needs of this vulnerable population.
The study, “Validation study of health administrative data algorithms to identify individuals experiencing homelessness and estimate population prevalence of homelessness in Ontario, Canada” is published in BMJ Open.